forked from ~ljy/RK356X_SDK_RELEASE
blame | history | patch | commit | commitdiff
hc
2023-12-11 072de836f53be56a70cecf70b43ae43b7ce17376 
 kernel/drivers/nvdimm/pfn_devs.c
....@@ -1,14 +1,6 @@
1
+// SPDX-License-Identifier: GPL-2.0-only
12 /*
23  * Copyright(c) 2013-2016 Intel Corporation. All rights reserved.
3
- *
4
- * This program is free software; you can redistribute it and/or modify
5
- * it under the terms of version 2 of the GNU General Public License as
6
- * published by the Free Software Foundation.
7
- *
8
- * This program is distributed in the hope that it will be useful, but
9
- * WITHOUT ANY WARRANTY; without even the implied warranty of
10
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
11
- * General Public License for more details.
124  */
135 #include <linux/memremap.h>
146 #include <linux/blkdev.h>
....@@ -33,17 +25,6 @@
3325 	kfree(nd_pfn->uuid);
3426 	kfree(nd_pfn);
3527 }
36
-
37
-static struct device_type nd_pfn_device_type = {
38
-	.name = "nd_pfn",
39
-	.release = nd_pfn_release,
40
-};
41
-
42
-bool is_nd_pfn(struct device *dev)
43
-{
44
-	return dev ? dev->type == &nd_pfn_device_type : false;
45
-}
46
-EXPORT_SYMBOL(is_nd_pfn);
4728 
4829 struct nd_pfn *to_nd_pfn(struct device *dev)
4930 {
....@@ -75,7 +56,7 @@
7556 	struct nd_pfn *nd_pfn = to_nd_pfn_safe(dev);
7657 	ssize_t rc = 0;
7758 
78
-	device_lock(dev);
59
+	nd_device_lock(dev);
7960 	nvdimm_bus_lock(dev);
8061 	if (dev->driver)
8162 		rc = -EBUSY;
....@@ -97,7 +78,7 @@
9778 	dev_dbg(dev, "result: %zd wrote: %s%s", rc, buf,
9879 			buf[len - 1] == '\n' ? "" : "\n");
9980 	nvdimm_bus_unlock(dev);
100
-	device_unlock(dev);
81
+	nd_device_unlock(dev);
10182 
10283 	return rc ? rc : len;
10384 }
....@@ -111,43 +92,46 @@
11192 	return sprintf(buf, "%ld\n", nd_pfn->align);
11293 }
11394 
114
-static const unsigned long *nd_pfn_supported_alignments(void)
95
+static unsigned long *nd_pfn_supported_alignments(unsigned long *alignments)
11596 {
116
-	/*
117
-	 * This needs to be a non-static variable because the *_SIZE
118
-	 * macros aren't always constants.
119
-	 */
120
-	const unsigned long supported_alignments[] = {
121
-		PAGE_SIZE,
122
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
123
-		HPAGE_PMD_SIZE,
124
-#ifdef CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
125
-		HPAGE_PUD_SIZE,
126
-#endif
127
-#endif
128
-		0,
129
-	};
130
-	static unsigned long data[ARRAY_SIZE(supported_alignments)];
13197 
132
-	memcpy(data, supported_alignments, sizeof(data));
98
+	alignments[0] = PAGE_SIZE;
13399 
134
-	return data;
100
+	if (has_transparent_hugepage()) {
101
+		alignments[1] = HPAGE_PMD_SIZE;
102
+		if (IS_ENABLED(CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD))
103
+			alignments[2] = HPAGE_PUD_SIZE;
104
+	}
105
+
106
+	return alignments;
107
+}
108
+
109
+/*
110
+ * Use pmd mapping if supported as default alignment
111
+ */
112
+static unsigned long nd_pfn_default_alignment(void)
113
+{
114
+
115
+	if (has_transparent_hugepage())
116
+		return HPAGE_PMD_SIZE;
117
+	return PAGE_SIZE;
135118 }
136119 
137120 static ssize_t align_store(struct device *dev,
138121 		struct device_attribute *attr, const char *buf, size_t len)
139122 {
140123 	struct nd_pfn *nd_pfn = to_nd_pfn_safe(dev);
124
+	unsigned long aligns[MAX_NVDIMM_ALIGN] = { [0] = 0, };
141125 	ssize_t rc;
142126 
143
-	device_lock(dev);
127
+	nd_device_lock(dev);
144128 	nvdimm_bus_lock(dev);
145129 	rc = nd_size_select_store(dev, buf, &nd_pfn->align,
146
-			nd_pfn_supported_alignments());
130
+			nd_pfn_supported_alignments(aligns));
147131 	dev_dbg(dev, "result: %zd wrote: %s%s", rc, buf,
148132 			buf[len - 1] == '\n' ? "" : "\n");
149133 	nvdimm_bus_unlock(dev);
150
-	device_unlock(dev);
134
+	nd_device_unlock(dev);
151135 
152136 	return rc ? rc : len;
153137 }
....@@ -169,11 +153,11 @@
169153 	struct nd_pfn *nd_pfn = to_nd_pfn_safe(dev);
170154 	ssize_t rc;
171155 
172
-	device_lock(dev);
156
+	nd_device_lock(dev);
173157 	rc = nd_uuid_store(dev, &nd_pfn->uuid, buf, len);
174158 	dev_dbg(dev, "result: %zd wrote: %s%s", rc, buf,
175159 			buf[len - 1] == '\n' ? "" : "\n");
176
-	device_unlock(dev);
160
+	nd_device_unlock(dev);
177161 
178162 	return rc ? rc : len;
179163 }
....@@ -198,13 +182,13 @@
198182 	struct nd_pfn *nd_pfn = to_nd_pfn_safe(dev);
199183 	ssize_t rc;
200184 
201
-	device_lock(dev);
185
+	nd_device_lock(dev);
202186 	nvdimm_bus_lock(dev);
203187 	rc = nd_namespace_store(dev, &nd_pfn->ndns, buf, len);
204188 	dev_dbg(dev, "result: %zd wrote: %s%s", rc, buf,
205189 			buf[len - 1] == '\n' ? "" : "\n");
206190 	nvdimm_bus_unlock(dev);
207
-	device_unlock(dev);
191
+	nd_device_unlock(dev);
208192 
209193 	return rc;
210194 }
....@@ -216,7 +200,7 @@
216200 	struct nd_pfn *nd_pfn = to_nd_pfn_safe(dev);
217201 	ssize_t rc;
218202 
219
-	device_lock(dev);
203
+	nd_device_lock(dev);
220204 	if (dev->driver) {
221205 		struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
222206 		u64 offset = __le64_to_cpu(pfn_sb->dataoff);
....@@ -230,11 +214,11 @@
230214 		/* no address to convey if the pfn instance is disabled */
231215 		rc = -ENXIO;
232216 	}
233
-	device_unlock(dev);
217
+	nd_device_unlock(dev);
234218 
235219 	return rc;
236220 }
237
-static DEVICE_ATTR_RO(resource);
221
+static DEVICE_ATTR_ADMIN_RO(resource);
238222 
239223 static ssize_t size_show(struct device *dev,
240224 		struct device_attribute *attr, char *buf)
....@@ -242,7 +226,7 @@
242226 	struct nd_pfn *nd_pfn = to_nd_pfn_safe(dev);
243227 	ssize_t rc;
244228 
245
-	device_lock(dev);
229
+	nd_device_lock(dev);
246230 	if (dev->driver) {
247231 		struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
248232 		u64 offset = __le64_to_cpu(pfn_sb->dataoff);
....@@ -258,7 +242,7 @@
258242 		/* no size to convey if the pfn instance is disabled */
259243 		rc = -ENXIO;
260244 	}
261
-	device_unlock(dev);
245
+	nd_device_unlock(dev);
262246 
263247 	return rc;
264248 }
....@@ -267,7 +251,10 @@
267251 static ssize_t supported_alignments_show(struct device *dev,
268252 		struct device_attribute *attr, char *buf)
269253 {
270
-	return nd_size_select_show(0, nd_pfn_supported_alignments(), buf);
254
+	unsigned long aligns[MAX_NVDIMM_ALIGN] = { [0] = 0, };
255
+
256
+	return nd_size_select_show(0,
257
+			nd_pfn_supported_alignments(aligns), buf);
271258 }
272259 static DEVICE_ATTR_RO(supported_alignments);
273260 
....@@ -282,24 +269,28 @@
282269 	NULL,
283270 };
284271 
285
-static umode_t pfn_visible(struct kobject *kobj, struct attribute *a, int n)
286
-{
287
-	if (a == &dev_attr_resource.attr)
288
-		return 0400;
289
-	return a->mode;
290
-}
291
-
292
-struct attribute_group nd_pfn_attribute_group = {
272
+static struct attribute_group nd_pfn_attribute_group = {
293273 	.attrs = nd_pfn_attributes,
294
-	.is_visible = pfn_visible,
295274 };
296275 
297
-static const struct attribute_group *nd_pfn_attribute_groups[] = {
276
+const struct attribute_group *nd_pfn_attribute_groups[] = {
298277 	&nd_pfn_attribute_group,
299278 	&nd_device_attribute_group,
300279 	&nd_numa_attribute_group,
301280 	NULL,
302281 };
282
+
283
+static const struct device_type nd_pfn_device_type = {
284
+	.name = "nd_pfn",
285
+	.release = nd_pfn_release,
286
+	.groups = nd_pfn_attribute_groups,
287
+};
288
+
289
+bool is_nd_pfn(struct device *dev)
290
+{
291
+	return dev ? dev->type == &nd_pfn_device_type : false;
292
+}
293
+EXPORT_SYMBOL(is_nd_pfn);
303294 
304295 struct device *nd_pfn_devinit(struct nd_pfn *nd_pfn,
305296 		struct nd_namespace_common *ndns)
....@@ -310,7 +301,7 @@
310301 		return NULL;
311302 
312303 	nd_pfn->mode = PFN_MODE_NONE;
313
-	nd_pfn->align = PFN_DEFAULT_ALIGNMENT;
304
+	nd_pfn->align = nd_pfn_default_alignment();
314305 	dev = &nd_pfn->dev;
315306 	device_initialize(&nd_pfn->dev);
316307 	if (ndns && !__nd_attach_ndns(&nd_pfn->dev, ndns, &nd_pfn->ndns)) {
....@@ -339,7 +330,6 @@
339330 
340331 	dev = &nd_pfn->dev;
341332 	dev_set_name(dev, "pfn%d.%d", nd_region->id, nd_pfn->id);
342
-	dev->groups = nd_pfn_attribute_groups;
343333 	dev->type = &nd_pfn_device_type;
344334 	dev->parent = &nd_region->dev;
345335 
....@@ -361,6 +351,89 @@
361351 	return dev;
362352 }
363353 
354
+/*
355
+ * nd_pfn_clear_memmap_errors() clears any errors in the volatile memmap
356
+ * space associated with the namespace. If the memmap is set to DRAM, then
357
+ * this is a no-op. Since the memmap area is freshly initialized during
358
+ * probe, we have an opportunity to clear any badblocks in this area.
359
+ */
360
+static int nd_pfn_clear_memmap_errors(struct nd_pfn *nd_pfn)
361
+{
362
+	struct nd_region *nd_region = to_nd_region(nd_pfn->dev.parent);
363
+	struct nd_namespace_common *ndns = nd_pfn->ndns;
364
+	void *zero_page = page_address(ZERO_PAGE(0));
365
+	struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
366
+	int num_bad, meta_num, rc, bb_present;
367
+	sector_t first_bad, meta_start;
368
+	struct nd_namespace_io *nsio;
369
+
370
+	if (nd_pfn->mode != PFN_MODE_PMEM)
371
+		return 0;
372
+
373
+	nsio = to_nd_namespace_io(&ndns->dev);
374
+	meta_start = (SZ_4K + sizeof(*pfn_sb)) >> 9;
375
+	meta_num = (le64_to_cpu(pfn_sb->dataoff) >> 9) - meta_start;
376
+
377
+	/*
378
+	 * re-enable the namespace with correct size so that we can access
379
+	 * the device memmap area.
380
+	 */
381
+	devm_namespace_disable(&nd_pfn->dev, ndns);
382
+	rc = devm_namespace_enable(&nd_pfn->dev, ndns, le64_to_cpu(pfn_sb->dataoff));
383
+	if (rc)
384
+		return rc;
385
+
386
+	do {
387
+		unsigned long zero_len;
388
+		u64 nsoff;
389
+
390
+		bb_present = badblocks_check(&nd_region->bb, meta_start,
391
+				meta_num, &first_bad, &num_bad);
392
+		if (bb_present) {
393
+			dev_dbg(&nd_pfn->dev, "meta: %x badblocks at %llx\n",
394
+					num_bad, first_bad);
395
+			nsoff = ALIGN_DOWN((nd_region->ndr_start
396
+					+ (first_bad << 9)) - nsio->res.start,
397
+					PAGE_SIZE);
398
+			zero_len = ALIGN(num_bad << 9, PAGE_SIZE);
399
+			while (zero_len) {
400
+				unsigned long chunk = min(zero_len, PAGE_SIZE);
401
+
402
+				rc = nvdimm_write_bytes(ndns, nsoff, zero_page,
403
+							chunk, 0);
404
+				if (rc)
405
+					break;
406
+
407
+				zero_len -= chunk;
408
+				nsoff += chunk;
409
+			}
410
+			if (rc) {
411
+				dev_err(&nd_pfn->dev,
412
+					"error clearing %x badblocks at %llx\n",
413
+					num_bad, first_bad);
414
+				return rc;
415
+			}
416
+		}
417
+	} while (bb_present);
418
+
419
+	return 0;
420
+}
421
+
422
+static bool nd_supported_alignment(unsigned long align)
423
+{
424
+	int i;
425
+	unsigned long supported[MAX_NVDIMM_ALIGN] = { [0] = 0, };
426
+
427
+	if (align == 0)
428
+		return false;
429
+
430
+	nd_pfn_supported_alignments(supported);
431
+	for (i = 0; supported[i]; i++)
432
+		if (align == supported[i])
433
+			return true;
434
+	return false;
435
+}
436
+
364437 /**
365438  * nd_pfn_validate - read and validate info-block
366439  * @nd_pfn: fsdax namespace runtime state / properties
....@@ -373,6 +446,7 @@
373446 int nd_pfn_validate(struct nd_pfn *nd_pfn, const char *sig)
374447 {
375448 	u64 checksum, offset;
449
+	struct resource *res;
376450 	enum nd_pfn_mode mode;
377451 	struct nd_namespace_io *nsio;
378452 	unsigned long align, start_pad;
....@@ -409,6 +483,11 @@
409483 	if (__le16_to_cpu(pfn_sb->version_minor) < 2)
410484 		pfn_sb->align = 0;
411485 
486
+	if (__le16_to_cpu(pfn_sb->version_minor) < 4) {
487
+		pfn_sb->page_struct_size = cpu_to_le16(64);
488
+		pfn_sb->page_size = cpu_to_le32(PAGE_SIZE);
489
+	}
490
+
412491 	switch (le32_to_cpu(pfn_sb->mode)) {
413492 	case PFN_MODE_RAM:
414493 	case PFN_MODE_PMEM:
....@@ -423,6 +502,34 @@
423502 	if (align == 0)
424503 		align = 1UL << ilog2(offset);
425504 	mode = le32_to_cpu(pfn_sb->mode);
505
+
506
+	if ((le32_to_cpu(pfn_sb->page_size) > PAGE_SIZE) &&
507
+			(mode == PFN_MODE_PMEM)) {
508
+		dev_err(&nd_pfn->dev,
509
+				"init failed, page size mismatch %d\n",
510
+				le32_to_cpu(pfn_sb->page_size));
511
+		return -EOPNOTSUPP;
512
+	}
513
+
514
+	if ((le16_to_cpu(pfn_sb->page_struct_size) < sizeof(struct page)) &&
515
+			(mode == PFN_MODE_PMEM)) {
516
+		dev_err(&nd_pfn->dev,
517
+				"init failed, struct page size mismatch %d\n",
518
+				le16_to_cpu(pfn_sb->page_struct_size));
519
+		return -EOPNOTSUPP;
520
+	}
521
+
522
+	/*
523
+	 * Check whether the we support the alignment. For Dax if the
524
+	 * superblock alignment is not matching, we won't initialize
525
+	 * the device.
526
+	 */
527
+	if (!nd_supported_alignment(align) &&
528
+			!memcmp(pfn_sb->signature, DAX_SIG, PFN_SIG_LEN)) {
529
+		dev_err(&nd_pfn->dev, "init failed, alignment mismatch: "
530
+				"%ld:%ld\n", nd_pfn->align, align);
531
+		return -EOPNOTSUPP;
532
+	}
426533 
427534 	if (!nd_pfn->uuid) {
428535 		/*
....@@ -455,14 +562,14 @@
455562 			dev_dbg(&nd_pfn->dev, "align: %lx:%lx mode: %d:%d\n",
456563 					nd_pfn->align, align, nd_pfn->mode,
457564 					mode);
458
-			return -EINVAL;
565
+			return -EOPNOTSUPP;
459566 		}
460567 	}
461568 
462569 	if (align > nvdimm_namespace_capacity(ndns)) {
463570 		dev_err(&nd_pfn->dev, "alignment: %lx exceeds capacity %llx\n",
464571 				align, nvdimm_namespace_capacity(ndns));
465
-		return -EINVAL;
572
+		return -EOPNOTSUPP;
466573 	}
467574 
468575 	/*
....@@ -472,18 +579,31 @@
472579 	 * established.
473580 	 */
474581 	nsio = to_nd_namespace_io(&ndns->dev);
475
-	if (offset >= resource_size(&nsio->res)) {
582
+	res = &nsio->res;
583
+	if (offset >= resource_size(res)) {
476584 		dev_err(&nd_pfn->dev, "pfn array size exceeds capacity of %s\n",
477585 				dev_name(&ndns->dev));
478
-		return -EBUSY;
586
+		return -EOPNOTSUPP;
479587 	}
480588 
481
-	if ((align && !IS_ALIGNED(nsio->res.start + offset + start_pad, align))
589
+	if ((align && !IS_ALIGNED(res->start + offset + start_pad, align))
482590 			|| !IS_ALIGNED(offset, PAGE_SIZE)) {
483591 		dev_err(&nd_pfn->dev,
484592 				"bad offset: %#llx dax disabled align: %#lx\n",
485593 				offset, align);
486
-		return -ENXIO;
594
+		return -EOPNOTSUPP;
595
+	}
596
+
597
+	if (!IS_ALIGNED(res->start + le32_to_cpu(pfn_sb->start_pad),
598
+				memremap_compat_align())) {
599
+		dev_err(&nd_pfn->dev, "resource start misaligned\n");
600
+		return -EOPNOTSUPP;
601
+	}
602
+
603
+	if (!IS_ALIGNED(res->end + 1 - le32_to_cpu(pfn_sb->end_trunc),
604
+				memremap_compat_align())) {
605
+		dev_err(&nd_pfn->dev, "resource end misaligned\n");
606
+		return -EOPNOTSUPP;
487607 	}
488608 
489609 	return 0;
....@@ -531,117 +651,79 @@
531651 EXPORT_SYMBOL(nd_pfn_probe);
532652 
533653 /*
534
- * We hotplug memory at section granularity, pad the reserved area from
535
- * the previous section base to the namespace base address.
654
+ * We hotplug memory at sub-section granularity, pad the reserved area
655
+ * from the previous section base to the namespace base address.
536656  */
537657 static unsigned long init_altmap_base(resource_size_t base)
538658 {
539659 	unsigned long base_pfn = PHYS_PFN(base);
540660 
541
-	return PFN_SECTION_ALIGN_DOWN(base_pfn);
661
+	return SUBSECTION_ALIGN_DOWN(base_pfn);
542662 }
543663 
544664 static unsigned long init_altmap_reserve(resource_size_t base)
545665 {
546
-	unsigned long reserve = PFN_UP(SZ_8K);
666
+	unsigned long reserve = nd_info_block_reserve() >> PAGE_SHIFT;
547667 	unsigned long base_pfn = PHYS_PFN(base);
548668 
549
-	reserve += base_pfn - PFN_SECTION_ALIGN_DOWN(base_pfn);
669
+	reserve += base_pfn - SUBSECTION_ALIGN_DOWN(base_pfn);
550670 	return reserve;
551671 }
552672 
553673 static int __nvdimm_setup_pfn(struct nd_pfn *nd_pfn, struct dev_pagemap *pgmap)
554674 {
555
-	struct resource *res = &pgmap->res;
675
+	struct range *range = &pgmap->range;
556676 	struct vmem_altmap *altmap = &pgmap->altmap;
557677 	struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
558678 	u64 offset = le64_to_cpu(pfn_sb->dataoff);
559679 	u32 start_pad = __le32_to_cpu(pfn_sb->start_pad);
560680 	u32 end_trunc = __le32_to_cpu(pfn_sb->end_trunc);
681
+	u32 reserve = nd_info_block_reserve();
561682 	struct nd_namespace_common *ndns = nd_pfn->ndns;
562683 	struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);
563684 	resource_size_t base = nsio->res.start + start_pad;
685
+	resource_size_t end = nsio->res.end - end_trunc;
564686 	struct vmem_altmap __altmap = {
565687 		.base_pfn = init_altmap_base(base),
566688 		.reserve = init_altmap_reserve(base),
689
+		.end_pfn = PHYS_PFN(end),
567690 	};
568691 
569
-	memcpy(res, &nsio->res, sizeof(*res));
570
-	res->start += start_pad;
571
-	res->end -= end_trunc;
572
-
692
+	*range = (struct range) {
693
+		.start = nsio->res.start + start_pad,
694
+		.end = nsio->res.end - end_trunc,
695
+	};
696
+	pgmap->nr_range = 1;
573697 	if (nd_pfn->mode == PFN_MODE_RAM) {
574
-		if (offset < SZ_8K)
698
+		if (offset < reserve)
575699 			return -EINVAL;
576700 		nd_pfn->npfns = le64_to_cpu(pfn_sb->npfns);
577
-		pgmap->altmap_valid = false;
578701 	} else if (nd_pfn->mode == PFN_MODE_PMEM) {
579
-		nd_pfn->npfns = PFN_SECTION_ALIGN_UP((resource_size(res)
580
-					- offset) / PAGE_SIZE);
702
+		nd_pfn->npfns = PHYS_PFN((range_len(range) - offset));
581703 		if (le64_to_cpu(nd_pfn->pfn_sb->npfns) > nd_pfn->npfns)
582704 			dev_info(&nd_pfn->dev,
583705 					"number of pfns truncated from %lld to %ld\n",
584706 					le64_to_cpu(nd_pfn->pfn_sb->npfns),
585707 					nd_pfn->npfns);
586708 		memcpy(altmap, &__altmap, sizeof(*altmap));
587
-		altmap->free = PHYS_PFN(offset - SZ_8K);
709
+		altmap->free = PHYS_PFN(offset - reserve);
588710 		altmap->alloc = 0;
589
-		pgmap->altmap_valid = true;
711
+		pgmap->flags |= PGMAP_ALTMAP_VALID;
590712 	} else
591713 		return -ENXIO;
592714 
593715 	return 0;
594716 }
595717 
596
-static u64 phys_pmem_align_down(struct nd_pfn *nd_pfn, u64 phys)
597
-{
598
-	return min_t(u64, PHYS_SECTION_ALIGN_DOWN(phys),
599
-			ALIGN_DOWN(phys, nd_pfn->align));
600
-}
601
-
602
-/*
603
- * Check if pmem collides with 'System RAM', or other regions when
604
- * section aligned.  Trim it accordingly.
605
- */
606
-static void trim_pfn_device(struct nd_pfn *nd_pfn, u32 *start_pad, u32 *end_trunc)
607
-{
608
-	struct nd_namespace_common *ndns = nd_pfn->ndns;
609
-	struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);
610
-	struct nd_region *nd_region = to_nd_region(nd_pfn->dev.parent);
611
-	const resource_size_t start = nsio->res.start;
612
-	const resource_size_t end = start + resource_size(&nsio->res);
613
-	resource_size_t adjust, size;
614
-
615
-	*start_pad = 0;
616
-	*end_trunc = 0;
617
-
618
-	adjust = start - PHYS_SECTION_ALIGN_DOWN(start);
619
-	size = resource_size(&nsio->res) + adjust;
620
-	if (region_intersects(start - adjust, size, IORESOURCE_SYSTEM_RAM,
621
-				IORES_DESC_NONE) == REGION_MIXED
622
-			|| nd_region_conflict(nd_region, start - adjust, size))
623
-		*start_pad = PHYS_SECTION_ALIGN_UP(start) - start;
624
-
625
-	/* Now check that end of the range does not collide. */
626
-	adjust = PHYS_SECTION_ALIGN_UP(end) - end;
627
-	size = resource_size(&nsio->res) + adjust;
628
-	if (region_intersects(start, size, IORESOURCE_SYSTEM_RAM,
629
-				IORES_DESC_NONE) == REGION_MIXED
630
-			|| !IS_ALIGNED(end, nd_pfn->align)
631
-			|| nd_region_conflict(nd_region, start, size))
632
-		*end_trunc = end - phys_pmem_align_down(nd_pfn, end);
633
-}
634
-
635718 static int nd_pfn_init(struct nd_pfn *nd_pfn)
636719 {
637
-	u32 dax_label_reserve = is_nd_dax(&nd_pfn->dev) ? SZ_128K : 0;
638720 	struct nd_namespace_common *ndns = nd_pfn->ndns;
639721 	struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);
640722 	resource_size_t start, size;
641723 	struct nd_region *nd_region;
642
-	u32 start_pad, end_trunc;
724
+	unsigned long npfns, align;
725
+	u32 end_trunc;
643726 	struct nd_pfn_sb *pfn_sb;
644
-	unsigned long npfns;
645727 	phys_addr_t offset;
646728 	const char *sig;
647729 	u64 checksum;
....@@ -658,6 +740,8 @@
658740 		sig = PFN_SIG;
659741 
660742 	rc = nd_pfn_validate(nd_pfn, sig);
743
+	if (rc == 0)
744
+		return nd_pfn_clear_memmap_errors(nd_pfn);
661745 	if (rc != -ENODEV)
662746 		return rc;
663747 
....@@ -672,44 +756,56 @@
672756 		return -ENXIO;
673757 	}
674758 
675
-	memset(pfn_sb, 0, sizeof(*pfn_sb));
676
-
677
-	trim_pfn_device(nd_pfn, &start_pad, &end_trunc);
678
-	if (start_pad + end_trunc)
679
-		dev_info(&nd_pfn->dev, "%s alignment collision, truncate %d bytes\n",
680
-				dev_name(&ndns->dev), start_pad + end_trunc);
681
-
682759 	/*
683760 	 * Note, we use 64 here for the standard size of struct page,
684761 	 * debugging options may cause it to be larger in which case the
685762 	 * implementation will limit the pfns advertised through
686763 	 * ->direct_access() to those that are included in the memmap.
687764 	 */
688
-	start = nsio->res.start + start_pad;
765
+	start = nsio->res.start;
689766 	size = resource_size(&nsio->res);
690
-	npfns = PFN_SECTION_ALIGN_UP((size - start_pad - end_trunc - SZ_8K)
691
-			/ PAGE_SIZE);
767
+	npfns = PHYS_PFN(size - SZ_8K);
768
+	align = max(nd_pfn->align, memremap_compat_align());
769
+
770
+	/*
771
+	 * When @start is misaligned fail namespace creation. See
772
+	 * the 'struct nd_pfn_sb' commentary on why ->start_pad is not
773
+	 * an option.
774
+	 */
775
+	if (!IS_ALIGNED(start, memremap_compat_align())) {
776
+		dev_err(&nd_pfn->dev, "%s: start %pa misaligned to %#lx\n",
777
+				dev_name(&ndns->dev), &start,
778
+				memremap_compat_align());
779
+		return -EINVAL;
780
+	}
781
+	end_trunc = start + size - ALIGN_DOWN(start + size, align);
692782 	if (nd_pfn->mode == PFN_MODE_PMEM) {
693783 		/*
694784 		 * The altmap should be padded out to the block size used
695785 		 * when populating the vmemmap. This *should* be equal to
696786 		 * PMD_SIZE for most architectures.
787
+		 *
788
+		 * Also make sure size of struct page is less than 64. We
789
+		 * want to make sure we use large enough size here so that
790
+		 * we don't have a dynamic reserve space depending on
791
+		 * struct page size. But we also want to make sure we notice
792
+		 * when we end up adding new elements to struct page.
697793 		 */
698
-		offset = ALIGN(start + SZ_8K + 64 * npfns + dax_label_reserve,
699
-				max(nd_pfn->align, PMD_SIZE)) - start;
794
+		BUILD_BUG_ON(sizeof(struct page) > MAX_STRUCT_PAGE_SIZE);
795
+		offset = ALIGN(start + SZ_8K + MAX_STRUCT_PAGE_SIZE * npfns, align)
796
+			- start;
700797 	} else if (nd_pfn->mode == PFN_MODE_RAM)
701
-		offset = ALIGN(start + SZ_8K + dax_label_reserve,
702
-				nd_pfn->align) - start;
798
+		offset = ALIGN(start + SZ_8K, align) - start;
703799 	else
704800 		return -ENXIO;
705801 
706
-	if (offset + start_pad + end_trunc >= size) {
802
+	if (offset >= size) {
707803 		dev_err(&nd_pfn->dev, "%s unable to satisfy requested alignment\n",
708804 				dev_name(&ndns->dev));
709805 		return -ENXIO;
710806 	}
711807 
712
-	npfns = (size - offset - start_pad - end_trunc) / SZ_4K;
808
+	npfns = PHYS_PFN(size - offset - end_trunc);
713809 	pfn_sb->mode = cpu_to_le32(nd_pfn->mode);
714810 	pfn_sb->dataoff = cpu_to_le64(offset);
715811 	pfn_sb->npfns = cpu_to_le64(npfns);
....@@ -717,13 +813,18 @@
717813 	memcpy(pfn_sb->uuid, nd_pfn->uuid, 16);
718814 	memcpy(pfn_sb->parent_uuid, nd_dev_to_uuid(&ndns->dev), 16);
719815 	pfn_sb->version_major = cpu_to_le16(1);
720
-	pfn_sb->version_minor = cpu_to_le16(3);
721
-	pfn_sb->start_pad = cpu_to_le32(start_pad);
816
+	pfn_sb->version_minor = cpu_to_le16(4);
722817 	pfn_sb->end_trunc = cpu_to_le32(end_trunc);
723818 	pfn_sb->align = cpu_to_le32(nd_pfn->align);
819
+	pfn_sb->page_struct_size = cpu_to_le16(MAX_STRUCT_PAGE_SIZE);
820
+	pfn_sb->page_size = cpu_to_le32(PAGE_SIZE);
724821 	checksum = nd_sb_checksum((struct nd_gen_sb *) pfn_sb);
725822 	pfn_sb->checksum = cpu_to_le64(checksum);
726823 
824
+	rc = nd_pfn_clear_memmap_errors(nd_pfn);
825
+	if (rc)
826
+		return rc;
827
+
727828 	return nvdimm_write_bytes(ndns, SZ_4K, pfn_sb, sizeof(*pfn_sb), 0);
728829 }
729830